Co-optimized Locational Pricing of Energy in a Competitive Market
Project Proposal
Project Supervisor:
Dr. Bhupendra Bimal Chhetri, Kathmandu University Mr Roshan Bhattarai, Electrical Department, Kathmandu University External Guide: Mr Mahesh Acharya, Project Co-ordinator, Nepal Electricity Authority
Submitted by:
Hitendra Dev Shakya MEPE Student, Kathmandu University
Project Proposal :
MEPE Student : Hitendra Dev Shakya
1. Title of the Project : Co-optimized Locational Pricing of Energy in a competitive market 2. Objective of the project: To study the optimization model for a competitive market energy pricing which incorporates locational pricing of energy as well as reserve capacity.
3. Discussion of the Project: In Nepal, we have east-to-west elongated geography and hence, we are naturally prone to transmission congestion problems in the east-west transmission lines. It means that the pricing in a competitive market has to take care of the congestion by locational pricing. Additionally, the congestion means the reserve capacity for system stability is also regional and there is a component of reserve capacity pricing in the locational pricing. For security of system, it is necessary to have reserves in the system which are called spinning reserves expressed in terms of ramp rates, but other reserves such as 10-minute or 30-minute non-spinning reserves such as gas turbines, pump turbines or fast hydro-turbines. In a competitive market, price has to be paid not only for the Energy used but also the reserve capacity demanded. Hence, a model of pricing that co-optimizes these expected costs of energy and reserve is very useful. In a competitive market, it is assumed that generators offer energy and reserve and cost offer based on their own optimization processes but which may be even speculative and manipulative if it enjoys market power. In any case, the co-optimization model should provide (for the ISO) the best scheduling and pricing mechanism based not on judgment and experience but on mathematical basis. It is assumed that Nepalese power sector will soon be unbundled and there will be an Independent System Operator (ISO) which will receive the offers from the generators and prepare the scheduling based on security-constrained optimization. It will be a single settlement market clearing. Further studies can be done on further settlements and even more variables optimizations.
There will be a base situation where the schedule and prices are defined through the optimization. In case of contingencies, there will be optimized solution the Payment will be based on exactly which contingency the real-time system occurred. If there are no contingency, the payment will be made on the base conditions. Study of Probable models of Restructuring for Nepal and Energy Pricing within it The market model : The above model is not the end model in itself for competitive electricity market. The project will also study different models in brief, with the Co-optimized pricing model in perspective. Models of Power Exchange, Poolco and competitive retail sales will be discussed as it is found in the literatures. Besides, the suitability of these models in relation to present Nepalese power sector will also be discussed but purely on a judgmental basis. The objective of this discussion on market model will be just to provide a proper understanding of the market mechanisms for power. 4. The methodology: The Co-optimization formulation for integrated Energy and Reserve market. 1. The supply curves of suppliers consisting of price and generation limits and reserve ramping 2. System demand at any point of time. The demand may vary according to a estimated load curve and load forecast which will form the base. 3. The base case will be a combination of supply and demand where the system is stable at the optimum cost of energy. 4. The contingencies (n-1) of system where the reserve will come into play to save the system. eg. unit out or line out or load increase from the estimated one. Further StudyOut of the above set-up, the ISO has to find the optimum scheduling and pricing mechanism. Further study can include the ancillary services such as voltage support and the mechanism should include the value for these services. Three or more parameter optimization should be studied as a further study to this project. The Objective function: min Σ pk{Σ[CPi(Pik)+CRi(Rik)]}
P,R k=0 i=1 K I
C
Pi(Pik)= energy cost for operating generator i at output level Pik in the kth contingency.
C
Ri(Rik)= reserve cost for operating generator i having reserve capacity Rik in the kth contingency. pk = the probability of k th contingency. The System constraints are nodal power balancing during base and contingencies – |Sik|<= Sjmax power flow constraints Vjmin<=Vjk<=Vjmax voltage limits Pimin<=Pik<=Pimax generation limits 0<=Rik<=Rimax spinning reserve limits (ramping) Pik+ Rik<= P imax Total Unit Committed Capacity (TUCC) – TUCC of unit i in the kth contingency is defined as Gik= Pik+ Rik The Co-optimization contains K+1 Optimal power flows (OPF) coupled by reserve cost and energy generation.
fk = min Σ pk{Σ[CPi(Pik)+CRi(Rik)]}
P,R k=0 i=1
K
I
The nodal pricing α ij=∆Gimin -------∆Dj βij=∆Gimax ---------∆Dj
Dj = real load at bus j α ij = sensitivity of change of Gimin w.r.t.load change at bus j. βij = sensitivity of change of Gimax w.r.t.load change at bus j.
Doing the perturbed Co-optimization for one extra unit of load at bus j of the K+1 systems and finding the optimum costs f0 and f1 and the difference f1 - f0 is the nodal energy price. Extension of the Project: It is expected to try to include the third marketed product that comes from the same generator – reactive power and the var reserve. Any generator is a multi-commodity device. It supplies energy, reserve and VAR at different levels and all of which has market value. The three are related and can be coupled. S=P+jQ, SR= G+ H, H= Q+QR where S is the apparent power supplied by generator, SR is the total apparent power capacity, G is the available capacity and H is the available VAR capacity, QR being the reserve VAR. Extended Objective Function: This coupling of the three parameters has to be further studied for constructing a proper Objective Function . In a real power system, absorption of reactive power can be useful at certain conditions. Hence, constraints have to be more defined. Such as states of the system at locations where positive VAR is rewarded and where negative VAR is rewarded. Further the linear assumption of the offers may also be examined further to represent more real conditions. 5. Procedure: 1. Study the optimization procedures and find an algorithm to be implemented in Matlab for the Objective function optimization. Adopt existing algorithm if appropriate. (eg MINOS interfaced into Matlab or some other commercial optimization packages). Improve on the objective function during further studies if possible. 2.Prepare a test case based on IEEE 30 bus system and certain contingencies in the system. 3. Apply the optimization models to the test system. 4. Analyse the results. 5. Prepare a Integrated Nepal Power System (INPS) probable system and apply the model. Assume a realistic values for Energy Price and Reserve Price. Convert some of the fixed take or pay PPAs into slightly differentiated energy and reserve prices. NEA generators will have different energy cost and reserve costs.
6. Apply the optimization model and obtain pricing for different contingencies and analyse the economic efficiency of the market (cost of the power). 7. Study a different objective function and model of pricing if necessary or appropriate. For example, the model should provide enough incentive for Investment in Generation Capacity Increment as well as Transmission investment in congestion areas without losing the Market Efficiency. The model should also be able to accommodate distortions in the competitive market such as subsidized tariff to certain population or geography for socio-political reasons. 8. Related study: 1. Study of pricing of energy by NEA or other IPPs to export power to India in view of Indian Electricity Act of 2003, Availability Based Tariff for Off-Peak energy. 2. Study few actual cases of generation pricing models. 6. Conclusion: The optimization model if found satisfactory should be readily applicable for ISO which gets unbundled from NEA as per Proposed Ordinance. The model should indicate further refinements where applicable for further studies.